Thèse de doctorat
Résumé : This thesis contains three essays covering different topics in the field of statistics

and econometrics of survey data. Chapters one and two analyse two aspects

of the Survey of Professional Forecasters (SPF hereafter) dataset. This survey

provides a large information on macroeconomic expectations done by the professional

forecasters and offers an opportunity to exploit a rich information set.

But it poses a challenge on how to extract the relevant information in a proper

way. The last chapter addresses the issue of analyzing the opinions on the euro

reported in the Flash Eurobaromenter dataset.

The first chapter Measuring Uncertainty and Disagreement in the European

Survey of Professional Forecasters proposes a density forecast methodology based

on the piecewise linear approximation of the individual’s forecasting histograms,

to measure uncertainty and disagreement of the professional forecasters. Since

1960 with the introduction of the SPF in the US, it has been clear that they were a

useful source of information to address the issue on how to measure disagreement

and uncertainty, without relying on macroeconomic or time series models. Direct

measures of uncertainty are seldom available, whereas many surveys report point

forecasts from a number of individual respondents. There has been a long tradition

of using measures of the dispersion of individual respondents’ point forecasts

(disagreement or consensus) as proxies for uncertainty. Unlike other surveys, the

SPF represents an exception. It directly asks for the point forecast, and for the

probability distribution, in the form of histogram, associated with the macro variables

of interest. An important issue that should be considered concerns how to

approximate individual probability densities and get accurate individual results

for disagreement and uncertainty before computing the aggregate measures. In

contrast to Zarnowitz and Lambros (1987), and Giordani and Soderlind (2003) we

overcome the problem associated with distributional assumptions of probability

density forecasts by using a non parametric approach that, instead of assuming

a functional form for the individual probability law, approximates the histogram

by a piecewise linear function. In addition, and unlike earlier works that focus on

US data, we employ European data, considering gross domestic product (GDP),

inflation and unemployment.

The second chapter Optimal Combination of Survey Forecasts is based on

a joint work with Christine De Mol and Domenico Giannone. It proposes an

approach to optimally combine survey forecasts, exploiting the whole covariance

structure among forecasters. There is a vast literature on forecast combination

methods, advocating their usefulness both from the theoretical and empirical

points of view (see e.g. the recent review by Timmermann (2006)). Surprisingly,

it appears that simple methods tend to outperform more sophisticated ones, as

shown for example by Genre et al. (2010) on the combination of the forecasts in

the SPF conducted by the European Central Bank (ECB). The main conclusion of

several studies is that the simple equal-weighted average constitutes a benchmark

that is hard to improve upon. In contrast to a great part of the literature which

does not exploit the correlation among forecasters, we take into account the full

covariance structure and we determine the optimal weights for the combination

of point forecasts as the minimizers of the mean squared forecast error (MSFE),

under the constraint that these weights are nonnegative and sum to one. We

compare our combination scheme with other methodologies in terms of forecasting

performance. Results show that the proposed optimal combination scheme is an

appropriate methodology to combine survey forecasts.

The literature on point forecast combination has been widely developed, however

there are fewer studies analyzing the issue for combination density forecast.

We extend our work considering the density forecasts combination. Moving from

the main results presented in Hall and Mitchell (2007), we propose an iterative

algorithm for computing the density weights which maximize the average logarithmic

score over the sample period. The empirical application is made for the

European GDP and inflation forecasts. Results suggest that optimal weights,

obtained via an iterative algorithm outperform the equal-weighted used by the

ECB density combinations.

The third chapter entitled Opinion surveys on the euro: a multilevel multinomial

logistic analysis outlines the multilevel aspects related to public attitudes

toward the euro. This work was motivated by the on-going debate whether the

perception of the euro among European citizenships after ten years from its introduction

was positive or negative. The aim of this work is, therefore, to disentangle

the issue of public attitudes considering either individual socio-demographic characteristics

and macroeconomic features of each country, counting each of them

as two separate levels in a single analysis. Considering a hierarchical structure

represents an advantage as it models within-country as well as between-country

relations using a single analysis. The multilevel analysis allows the consideration

of the existence of dependence between individuals within countries induced by

unobserved heterogeneity between countries, i.e. we include in the estimation

specific country characteristics not directly observable. In this chapter we empirically

investigate which individual characteristics and country specificities are

most important and affect the perception of the euro. The attitudes toward the

euro vary across individuals and countries, and are driven by personal considerations

based on the benefits and costs of using the single currency. Individual

features, such as a high level of education or living in a metropolitan area, have

a positive impact on the perception of the euro. Moreover, the country-specific

economic condition can influence individuals attitudes.